INACTIVE - http://mzl.la/ghe-archive - Basic framework for rapid creation of web tools
Перейти к файлу
Mark Côté fbf5f59528 Version bump to 0.6.2. 2012-11-06 18:17:37 -05:00
scripts Can now specify name instead of having cwd be the project root for most 2012-01-10 23:28:17 -05:00
templeton Bug 809242 - json_response() dumps datetime objects as ISO-format strings. r=wlach 2012-11-06 18:13:55 -05:00
.gitignore Ignore .pyc files 2011-11-21 12:52:26 -05:00
README.md Do nonblocking reads from subprocesses. Added /app -> /app/ redirect to nginx config. 2012-01-06 12:25:40 -05:00
setup.py Version bump to 0.6.2. 2012-11-06 18:17:37 -05:00

README.md

templeton, a minimal web framework

Templeton is a script, daemon, module, and support files intended for rapid development of web applications.

It currently runs on Linux. Development mode should run on OS X, but I haven't tried deployment mode.

Templeton apps

Templeton apps are essentially web.py apps with various enhancements and a (highly) suggested structure. They are managed by the templeton app server, templetond.

The main goal of templeton is to streamline the process of developing and deploying web apps. While developing, templeton provides a simple development server and a module to handle a lot of the underlying mechanics of modern web apps, such as decoding URL search/query strings and serving simple Python objects as JSON. When the app is ready to deploy, the templeton app server manages the app process and creates appropriate web-server config files.

Pursuant to this goal, templeton preserves the relative URL path in development and deployment. The development server run at the root directory of localhost on a given port. When deployed, the app is configured to have the same structure but live in a subpath, alongside other applications and content.

Templeton favours highly dynamic web sites based on the Single-Page Application pattern. To this end, static files are served at the base of the application's path, and dynamic content is (by default) served under api/. It is expected that index.html would be a static file that, with the help of JavaScript, would load dynamic content asynchronously. At the same time, this dynamic data is available at a known place to external applications.

(There are plans to remove this restriction so that templeton apps can also follow a template-based pattern, with static content consisting mainly of images, CSS, and JS.)

Templeton also packages up some common support web files, such as the jQuery library. The development server makes these available under /templeton. For deployment, the templeton command 'install' will copy the templeton/ directory to the root of a web server's html dir, again, so the apps can link to these files at a known location valid in both development and deployment modes.

For deployment, templeton currently supports nginx with spawn-fastcgi.

The templeton script

The templeton script is the user's point of contact with the whole templeton system.

Usage:

templeton install <www-data-dir>

Copies support files (JS, CSS) into a "templeton" directory in www-data-dir. The latter should be the root of the web site that will serve templeton apps, since the template HTML file loads JS and CSS from /templeton.

templeton init <appname>

Creates a new templeton project in a directory called appname, relative to the current working directory. If virtualenv is available, a virtualenv is installed to the new directory, and the template files are installed to appname/src/appname/. If virtualenv is not available, the templates are copied directly into appname/.

You should be able to serve up your default app by doing

# virtualenv available
source <appname>/bin/activate
cd <appname>/src/<appname>/server
python server.py [port]  # [port] defaults to 8080

or

# no virtualenv
cd <appname>/server
python server.py [port]  # [port] defaults to 8080

Go to http://localhost:8080/ to see the result. The next steps you'll want to do is edit handlers.py and put in your server-side business logic, and edit and create the files in ../html/ to build up your client-side logic.

The rest of the commands interface with the templeton app server:

templeton register

Registers the app in the current working directory with the app server. This also creates a web-server config file, for deployment. Note that you (probably) have to reload or restart your web server in order to pick up this config.

templeton unregister

Removes this app from the care of the app server.

templeton disable

Tells the app server not to start this app automatically.

templeton enable

Tells the app server to start this app automatically, if it was previously disabled.

templeton start

Starts the FastCGI process for dynamic content.

templeton stop

Stops the FastCGI process.

templeton list

Lists all templeton apps, with their status.

The templeton module

The templeton module has two main functions:

  • set up middleware to separate static pages from dynamic calls.
  • provide helpers for common tasks, such as handling specific request types.

middleware

The middleware module provides the patch_middleware() function to patch web.py's development server to serve the app with the standard templeton structure. The server.py template includes the necessary call.

handlers

templeton is geared toward client-rich, REST-based web applications. These typically involve a large amount of JSON. templeton provides decorators to simplify handler code.

  • @json_response is a decorator function that expects the decorated function to return a JSON-serializable object, which it uses to construct a proper web.py response.

  • @png_response expects the decorated function to return a PIL object, which it serves as a PNG image.

Both decoraters also redirect any URLs that don't end in a slash to one that does, to enforce some consistency. Thus if you access api/foo, it will redirect to api/foo/. Your handler URLs should reflect this.

The handlers module also provides helper functions.

load_urls() takes a web.py URL-handler sequence, i.e. (path, class_name, path, class_name, ...), and prepends the REST API path, '/api', to each given path. The default server.py (created by the 'init' script command) uses this function to load urls from handlers.py.

get_request_parms() parses the current request's search string and body as JSON and returns the results as (args, body).

A trivial example of a JSON handler that echoes back any search-string args:

import templeton

urls = (
    '/test/?', 'JsonTest'
)

class JsonTest(object):

    @templeton.handlers.json_response
    def GET(self):
        args, body = templeton.handlers.get_request_parms()
        return args

With the development server running on port 8080,

$ curl 'http://localhost:8080/api/test/?cat=meow&dog=bark&cat=purr'
{"dog": ["bark"], "cat": ["meow", "purr"]}